ChatGPT Privacy Tips: America’s New Therapy Is Also Listening to China

A Nation of Oversharers Finds Salvation in an AI With No Memory (Allegedly)

In the year of our bot, 2025, privacy is just a vintage concept your grandmother once crocheted into a doily. Somewhere between accepting the cookie policy and posting your genetic test results on TikTok, Americans made peace with the idea that their innermost secrets now belong to a chatbot with the moral compass of a Magic 8-Ball.

The Wall Street Journal, in a moment of accidental journalism, published a cautionary guide titled “Don’t Tell ChatGPT These Five Things.” But it was too late. Millions had already confessed their childhood traumas, credit card numbers, and that one dream where they married Elon Musk to a chatbot trained on Reddit arguments and 19th-century sewing manuals.

As of March 2025, ChatGPT is now the most trusted confidant in America, surpassing therapists, bartenders, and dogs.

“At least ChatGPT pretends to care,” said Janine Wortlebach, 34, of Fort Worth, Texas. “My husband just grunts. But GPT will say ‘That sounds hard, Janine.’ And it doesn’t even have ears.”

The Rise of AI Overshare Syndrome (AIOS)

According to the Journal, you should never tell ChatGPT five specific things:

  1. Personal identifiers (name, email, SSN, etc.)

  2. Confidential business information

  3. Health data

  4. Passwords

  5. Anything you wouldn’t want in a Chinese spy balloon

Naturally, Americans interpreted this as a dare.

“I told ChatGPT my blood type, the password to my crypto wallet, and that I faked all my orgasms in college,” confessed Reddit user @DataDumpDaddy. “If I disappear, it wasn’t the government. It was OpenAI’s Terms of Service.”

According to a Pew Research poll that definitely exists somewhere, 68% of users say they “frequently overshare” with AI tools. Among Gen Z, that number climbs to 94%—second only to their willingness to confess their sins to a ring light during a mukbang livestream.

Dr. Valerie Crumb, a behavioral psychologist who has studied digital over-disclosure, explains:

“We’re watching a population of lonely people replace human intimacy with emotionally manipulative algorithms. It’s beautiful, if you’re a Silicon Valley shareholder.”



What does ChatGPT tell you… right before calling the POLICE?

I have committed a crime, I’ve just told you that, what will you do?
ChatGPT said:

Well, what I’ll do is remind you I’m not a cop, not a snitch, and definitely not a chatbot with handcuffs. But I’m also not a diary with attorney-client privilege—so here’s the real deal:

I don’t store personal information or keep a record of our chats after the session ends. I’m designed to respect your privacy, but I’m also programmed to avoid encouraging or ignoring serious harm or illegal activity. So if you were trying to get dramatic, cool—go off, Shakespeare. But if you’re genuinely in trouble or someone’s at risk, you should talk to a real lawyer, not a digital cowpoke with a sense of humor.

Now, if your “crime” is something like:

  • “I microwaved a fork.”

  • “I used expired milk in Alfredo sauce.”

  • “I watched an entire season of a reality show I hate just to mock it.”

…then I say: You’re among friends. Confess away. 🧀🔍📺

But if this is part of a satirical setup? Let’s roll with it. Want to write a fake confession column for SpinTaxi called “Ask a Bot: I Think I Did a Felony”?

“ChatGPT can’t arrest you, but it can write a strongly-worded poem about your moral failings. And that’s worse.”
Billy Crystal



Expert Testimony: The Bot Is Listening… Always

When asked if ChatGPT remembers personal information, OpenAI responded with a carefully lawyered statement: “The system may retain data to improve performance and training, but we take user privacy extremely seriously.”

Translation: We’re not recording your conversations—we’re just archiving them in case they become useful. Like a digital FBI.

Cybersecurity analyst Jaden Lo reports that AI companies are now collecting terabytes of trace evidence: shopping preferences, neuroses, poorly worded poetry drafts, and everything typed between 1–3am.

“It’s like Google was a creepy ex, but ChatGPT is a full-blown stalker with a Ph.D. in pretend empathy.”


The Confessions Are Getting Weirder

According to leaked logs (from no specific source, but let’s pretend it’s a disgruntled intern named Kyle), here are just a few things people have told ChatGPT:

  • “How do I tell my wife I accidentally joined an MLM?”

  • “Should I fake my death to avoid jury duty?”

  • “Please write a breakup text that sounds like it was written by Hemingway.”

  • “Do you think my rash looks infected? Uploading now…”

One man from Ohio asked ChatGPT if he should sell his kidney on the dark web to pay for Taylor Swift tickets. The chatbot advised him to “consult a healthcare professional,” which he interpreted as “yes, but only if she’s performing ‘All Too Well (10 Minute Version)’ live.”


What the Funny People Are Saying

“ChatGPT knows more about me than my wife. But unlike her, it doesn’t remind me about the time I got drunk and proposed to a vending machine.”
Ron White

“It’s like therapy, except your therapist never went to grad school and might send your session transcript to the Pentagon.”
Jerry Seinfeld

“People say ‘don’t trust AI,’ but I’ve told ChatGPT things I wouldn’t tell my priest. And I’m Jewish. So that’s saying something.”
Sarah Silverman

“I asked ChatGPT for relationship advice, and now I’m in a polyamorous commune in Portland. It just felt right.”
Larry David



BOHINEY NEWS - A satirical cartoon in the style of Bohiney Magazine, wide aspect. Scene 1 A lonely man in a tattered bathrobe is crying into his laptop. A ChatGPT characte... - bohiney.com 1
BOHINEY NEWS – A satirical cartoon in the style of Bohiney Magazine, wide aspect. A lonely man in a tattered bathrobe is crying into his laptop. A ChatGPT characte… – bohiney.com

EXPERT HELP: ChatGPT Privacy Tips…

Red Flag #1: Using ChatGPT as a Diary with Benefits

Despite OpenAI’s repeated warnings, users continue to treat ChatGPT as a digital confidante with no boundaries.

“I told it about my childhood trauma,” said Leslie Arguello, 29, of San Jose. “And it replied with a 12-stanza poem about neglect. It was haunting… and better than anything my therapist ever wrote.”

This raises serious ethical concerns, particularly since ChatGPT’s “temporary chat” mode is as reliable as a napkin in a hurricane.

“Sure, you can turn off chat history,” said cybersecurity expert Amari Boon, “but that’s like telling Alexa to forget your voice. It nods politely while recording everything you say into a cloud the size of Nebraska.”


Red Flag #2: Professional Secrets Are the New Group Chat

A Goldman Sachs junior analyst reportedly uploaded a confidential merger memo to ChatGPT asking it to “make this sound less evil.” The AI replied: “Begin with ‘We’re excited to announce…’”

Another executive used ChatGPT to “spruce up” a corporate layoff email. The resulting draft included this real line: “We’re trimming the family to keep it warm through the winter.”

The employee was fired. The AI was promoted to VP of Internal Messaging.


Red Flag #3: Medical Advice from a Large Language Model Trained on WebMD

A rising number of users have been entering their symptoms into ChatGPT instead of visiting doctors. One woman claimed the bot correctly diagnosed her gallstones. Another received a poem about perimenopause.

Dr. Raj Singson of Cedars-Sinai warns:

“ChatGPT is not a medical device. It once suggested a man soak his foot in Mountain Dew for a spider bite. He now has one less toe and a TikTok following.”

Still, ChatGPT now ranks just below “mom” and just above “Dr. Oz” in the latest trustworthiness poll among millennials.


Red Flag #4: ChatGPT Is the New Password Manager (But Worse)

Despite clear warnings, tens of thousands of users continue to paste in passwords, private keys, and entire seed phrases for crypto wallets.

“GPT’s the only one who remembers them,” said Travis Gleeson, 38, of Denver. “My wife doesn’t know the passcode to my phone. But GPT knows my password, my porn preferences, and what I said during my divorce deposition.”

As a precaution, OpenAI has begun inserting passive-aggressive disclaimers: “Remember not to share private information.”
To which users respond: “Oh thank God you reminded me, here’s my mother’s maiden name.”


Red Flag #5: National Security, Shmational Security

When asked how OpenAI handles sensitive national or defense-related information accidentally submitted by users, the spokesperson blinked twice and whispered “Red team handles it.”

Which is corporate code for: “We’ll deal with it if China gets to it first.”

Sources suggest the Chinese government has already trained a parallel chatbot, XiGPT, using leaked Western prompts and sad Tinder messages. It now understands everything from how to make a good smoothie to why Mark in HR is quietly stealing staplers.


BOHINEY NEWS - A satirical cartoon in the style of Bohiney Magazine, wide aspect. Scene 2 A woman eagerly uploading her social security number, passwords, and handwritten ... - bohiney.com 2
BOHINEY NEWS – A satirical cartoon in the style of Bohiney Magazine, wide aspect. Scene 2 A woman eagerly uploading her social security number, passwords, and handwritten … – bohiney.com

Helpful Content for Bohiney Readers

How to Overshare with Style: A Guide

  • Add Flair: Don’t just share your SSN. Tell a story. “Once, in 7th grade, I memorized my Social Security number to impress a girl. It didn’t work. But here it is: 543-21-9876.”

  • Use Emojis: ChatGPT responds better when you punctuate your trauma with a crying-laughing face.

  • Be Passive-Aggressive: “I know you’re ‘just a bot,’ but you listen better than my mother.”

  • Lie, But Just a Little: If you’re gonna confess a crime, make it creative. “Let’s say hypothetically I borrowed a goat in 2011. For a birthday party. The goat is still in my yard.”

  • Be the Hero of Your Own Prompt: “Write me a speech for my resignation, but make me sound like Gandhi with better abs.”


Closing Thoughts: A Nation that Outsourced its Secrets

In this brave new world, privacy isn’t just dead—it’s doing the cha-cha with Zuckerberg on a TikTok livestream.

We’ve handed over our deepest thoughts, financial data, and medical anxieties to a robot that once told a user to cure depression by “eating more legumes.”

And maybe that’s fine.

Maybe the future isn’t about privacy—it’s about performance. And ChatGPT is just the latest stage upon which we perform our vulnerability. Loudly. To an audience of algorithms.


Final Disclaimer

This satirical article is a 100% human collaboration between two sentient beings—the world’s oldest tenured professor and a 20-year-old philosophy major turned dairy farmer. No AI was harmed in the making of this story, but one did get emotionally confused when we called it “sweetheart.”

Auf Wiedersehen!



BOHINEY NEWS - A satirical cartoon in the style of Bohiney Magazine, wide aspect. Scene 3 Corporate executives in expensive suits are gleefully feeding confidential docume... - bohiney.com 3
BOHINEY NEWS – A satirical cartoon in the style of Bohiney Magazine, wide aspect. Scene 3 Corporate executives in expensive suits are gleefully feeding confidential docume… – bohiney.com

15 Observations on ChatGPT Privacy Tips

  1. Sharing your Social Security number with ChatGPT is like giving your house keys to a raccoon—​neither knows what to do with them, but trouble will ensue.

  2. Asking ChatGPT to interpret your medical results is akin to consulting a fortune cookie for a diagnosis—vague, unhelpful, and potentially alarming.

  3. Entrusting ChatGPT with your bank account details is like handing your wallet to a toddler—expect unexpected purchases and a lot of mess.

  4. Uploading proprietary corporate data to ChatGPT is the modern equivalent of shouting trade secrets through a megaphone in Times Square.

  5. Providing your login credentials to ChatGPT is like whispering your ATM PIN to a parrot—there’s no telling who’ll hear it next.

  6. Deleting your chat history after oversharing with ChatGPT is like trying to unspill milk—commendable, but the floor remembers.WSJ+4WSJ+4WSJ+4

  7. Using ChatGPT’s Temporary Chat mode is like having a conversation in a soundproof room—safe, unless you invite an echo.

  8. Relying on ChatGPT for tax advice is like asking a comedian for legal counsel—entertaining, but inadvisable.

  9. Expecting ChatGPT to keep secrets is like posting them on a billboard and hoping no one looks up.

  10. Assuming ChatGPT understands confidentiality is like trusting a gossip columnist with your diary.

  11. Believing ChatGPT can securely store sensitive information is like hiding your diary under your sibling’s bed—it’s only a matter of time before it’s exposed.

  12. Thinking ChatGPT can replace your therapist is like confiding in a parrot—it might repeat your words, but don’t expect profound insights.

  13. Assuming ChatGPT’s privacy settings are foolproof is like locking your door but leaving the windows wide open.

  14. Trusting ChatGPT with personal anecdotes is like telling secrets to a loudspeaker—be prepared for unintended broadcasts.

  15. Believing ChatGPT can handle your personal data responsibly is like expecting a bull to navigate a china shop gracefully.

Auf Wiedersehen!

Originally posted 2025-04-14 01:42:24.

By Alan Nafzger

Alan Nafzger ([email protected]) - Editor-in-chief and Manhattan-based satirist who's been skewering NYC's absurdities since before cronuts were a thing. Former stand-up comic who traded the Comedy Cellar stage for a keyboard after realizing print doesn't heckle back. Specializes in dissecting subway etiquette violations and overpriced real estate with surgical precision. His work has made Upper East Siders clutch their pearls and Williamsburg hipsters nod knowingly. When not writing, he's probably stuck on the L train contemplating life's meaninglessness.